Poor Performance of Drools 5.0.x

[thekepat.sudesh]

Hi All,

Firstly I must congratulate on a very well designed Drools Expert, flow & fusion product. What it falls short is in performance. For example, you may run the conway GUI in 2.5 drools & it gives almost 0 cpu load, while the same in 5.0.x gives over 1 cpu on full load (1/4 logical cpus, Q6600 2.8Ghz). I wanted to use the same for several projects of mine, but now think twice. The reteOO , I thought was perfected in 4.0.x, but then why this performance nightmare? Please Help, since I really do like this product, as it is designed well.

Any performance benchmarks with ESPER?


Thanks a lot,

Sudesh

[kconner]

Is this a general drools question or specific to the ESB integration?

If the latter can you expand on the problem?

Thanks

[burrsutter]

I asked Edson what he thought and here was his response:

Drools 3+ is a full capabilities rules engine, while Drools 2.x was a partial implementation (good for some cases, but not for the general case). It is hard to tell you "what is wrong" or what "performance problems" are you seeing without further details. What we do know is that we are in the right track for real use cases (non-hello world examples).

Some numbers, if you want:

* Drools 3 gives you 10 times the throughput of Drools 2.5:

http://blog.athico.com/2006/11/rush-hour-and-content-based-routing.html

* Drools 4 vs Drools 5 on banking performance benchmark:

http://blogs.illation.com.au/2009/07/drools-performance-comparison/

AFAIK, Drools offers you the most complete set of features with one of the best performance results you can get from a full rules engine in the market. Of course we can and will continue improving, but we need a concrete problem in order to offer a concrete solution.

Based on the way you phrased the initial posting, I believe you are better off putting this concern into the Drools IRC/email list.

Burr